QPInverseProblemSolver Question: Strange Result after adding Interative collision #180
Unanswered
TeoREN0217
asked this question in
Help - Programming with SoftRobots
Replies: 2 comments 1 reply
-
Hi~ @hugtalbot @EulalieCoevoet |
Beta Was this translation helpful? Give feedback.
0 replies
-
Hi @TeoREN0217, is it possible to test your scene? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi! I've been working with a trunk-like soft robot recently and I found
QPInverseProblemSolver
has very strange results. Here is the problem:Without collision in the environment, the tip has a normal result like:
But after adding interactive collision, the result always makes the robot tip lean to the left (in different youngmodules, positions, robot FEM mesh after I trying to find the problem but always left!) like:
I'm pretty sure it's not the environment collision model because I tried to cut different parts from the model and it didn't fix the problem. Also, there is a big force from on of the cable actuator makes the robot transform like:
I tried to add a constraint of X axis but it still make robot transform. The cable position is symmetrical and the No.1 cable actuator has bigger force than No.2 cable actuator because after animating a while, No.2 cable force will be 0 and only No.1 has force and that why the robot lean to left. These whole thing worked in Inverse mode so the
QPInverseProblemSolver
solved the value. I was wondering isQPInverseProblemSolver
has some bugs or something wrong with my model?Here is the displacement and force exported. For no matter the robot and it's actuator or the environment collision models are all symmetrical, I think the two cable should has a similar result.
Higher force make the displacement lower and make the robot transformed.
Beta Was this translation helpful? Give feedback.
All reactions